The proposed research is designed to investigate memory for "forgotten" items (i.e., memory for items below the performance threshold of a given test). These investigations primarily include studies of the theoretical and empirical factors underlying the feeling of knowing, both in terms of what it is based upon and in terms of its validity for predicting subsequent memory performance. Also included are studies of savings during relearning and studies of spontaneous recovery. Using the results of the above investigations and other related findings, an attempt is made to determine ways of improving memory retrieval for previously acquired information. The items used in this research come both from archival information acquired outside the laboratory and from information acquired in the laboratory.